PLOTTING IN R!

• Reproducibility
• Consistency
  
• Sharing code
• Many output types, high resolution

• “quick additions”
• A ton of text
• Conceptual figures

Illustrator

• Nice because figures are linked
• But expensive
• Bad for sharing

Powerpoint

• Everyone has it… Is that a good enough reason?

plot(sample(100),sample(100))

par(mar=c(3,3,1,1), mgp=c(1.5,0.5,0), tck=-0.02)
plot(sample(100),sample(100))

par(mar=c(3,3,1,1), mgp=c(1.5,0.5,0), tck=-0.02)
plot(sample(100), sample(100), pch=21, bg='red3', cex=2)

add.alpha <- function(col, alpha=1) {
 apply(sapply(col, col2rgb)/255, 2,
       function(x)
         rgb(x[1], x[2], x[3], alpha=alpha)) 
}

par(mar=c(3,3,1,1), mgp=c(1.5,0.5,0), tck=-0.02)
plot(sample(100), sample(100), pch=21, bg=add.alpha('red3',0.7), cex=2)

A few options in R

• par(mfrow = c(1,2))
  • very easy, Hilary uses 50% of the time
• layout(matrix(c(1,1,2,3), 2, 2, byrow = TRUE))
  • confusing, Hilary uses 0% of the time
• par(fig = c(0,0.8,0,0.8), new = TRUE)
  • harder, but useful, Hilary uses 50% of the time

par(mar=c(3,3,1,1), mgp=c(1.5,0.5,0), tck=-0.02, mfrow=c(2,1))
plot(sample(100), sample(100), pch=21, bg=add.alpha('red3',0.7))
plot(sample(100), sample(100), pch=21, bg=add.alpha('slateblue',0.7))

par(mar=c(3,3,1,1), mgp=c(1.5,0.5,0))
par(fig=c(0,0.6,0,0.6))
plot(sample(100), sample(100),  pch=21,bg=add.alpha('red3',0.7))
par(fig=c(0.6,1,0,1), new = TRUE)
plot(sample(100), sample(100), pch=21, bg=add.alpha('slateblue',0.7))

par(mar=c(3,3,1,1), mgp=c(1.5,0.5,0), tck=-0.02,ps=18)
plot(sample(100), sample(100), pch=21, bg=add.alpha('red3',0.7))
par(ps=14)
mtext(side = 3,line = 0,text = 'title')

Many figure formats

• jpeg (small)
• pdf(small, great resolution)
• tiff (huge)
• png (Hilary's favorite)

png('file.png')
...
dev.off()

png('Figures/file.png',height = 5,width = 8,units = 'in',res = 300)
  par(mar=c(3,3,1,1), mgp=c(1.5,0.5,0), tck=-0.02)
  plot(sample(100), sample(100), pch=21, bg=add.alpha('red3',0.7), cex=2)
dev.off()

quartz_off_screen 
                2 

png('Figures/file_poster.png',height = 5,width = 8,units = 'in',res = 300)
  par(mar=c(3,3,1,1), mgp=c(1.5,0.5,0), tck=-0.02, cex=2, bg='black',
    col.lab='grey90', col.axis='grey90', fg='grey90')
  plot(sample(100), sample(100), pch=21, bg=add.alpha('red3',0.7), cex=2)
  par(ps=12)
  mtext(side = 3, line = 0, text = 'title')
dev.off()

quartz_off_screen 
                2 

A consistent problem in collaborations

• Default fonts
  • Matlab - Helvetica
  • R - TT Courier New
  • Powerpoint - Calibri
• Margins
• Colors
• Sizes

Plot secchi depth of all lakes in the 2007 NLA dataset

secchi = read.csv('Data/NLA2007_Secchi_20091008.csv', stringsAsFactors = F)
loc = read.csv('Data/NLA2007_SampledLakeInformation_20091113.csv', stringsAsFactors = F)

library(dplyr)
combo = secchi %>% select(SITE_ID,SECMEAN) %>%
  group_by(SITE_ID) %>%
  summarise_each(funs(mean(., na.rm = TRUE))) %>%
  ungroup() %>%
  left_join(.,loc,by='SITE_ID') %>%
  select(SITE_ID, SECMEAN, LON_DD, LAT_DD, AREA_HA)

save(combo,file = "Data/combo.RData")

quantile(combo$SECMEAN,na.rm=T)

   0%   25%   50%   75%  100% 
 0.04  0.61  1.35  2.85 36.71 

quantile(combo$AREA_HA)

          0%          25%          50%          75%         100% 
4.036675e+00 2.424719e+01 6.982455e+01 2.847939e+02 1.674896e+05 

breaksArea = cut(combo$AREA_HA,breaks = c(1,10,100,1000,200000),labels = 1:4)
sizes = c(0.6,0.8,1,1.2)[breaksArea]
breaksSecchi = cut(combo$SECMEAN,breaks = c(0,1,5,10,50), labels = 1:4)
cols = c('moccasin','orange3','red4','black')[breaksSecchi]

library(maps)
par(mar=c(0,0,0,0))
map('state') #states maps

library(maps)
par(mar=c(0,0,0,0))
map('state', fill = T, border ='grey30', col='transparent', lwd=0.3, mar=c(0.5,0.5,0.5,3), projection = 'albers', par=c(20,40)) #states maps

library(maps);library(mapproj)
par(mar=c(0,0,0,0))
map('state', fill = T, border ='grey30', col='transparent', lwd=0.3,mar=c(0.5,0.5,0.5,3), projection = 'albers', par=c(20,40)) #states maps
points(mapproject(combo$LON_DD,combo$LAT_DD), bg=cols, lwd=0.5, col='black', pch=21, cex=sizes) # Add points with secchi data

  par(fig = c(0, 1, 0.1, 0.48)) # Add legend
  legend('bottomleft',legend=c(paste0('0-1n, (n=',table(breaksSecchi)[1],')'),
      paste0('1-5 m, (n=',table(breaksSecchi)[2],')'),
      paste0('5-10 m, (n=',table(breaksSecchi)[3],')'),
      paste0('> 10 m, (n=',table(breaksSecchi)[4],')')),
      pt.bg = c('moccasin','orange3','red4','black'),cex=1,bty='n',ncol=1,
      pch=21,pt.cex=1.2,col = 'black',title = 'Secchi Depth')